Cost of Equity (Ke)
How we compute the discount rate used in our DCF and dividend discount models. The number you see on a stock's valuation page is the number we actually apply — no behind-the-scenes smoothing.
The formula
We use the Capital Asset Pricing Model (CAPM):
Ke = Risk-Free Rate + β × Market Risk PremiumEach input is sourced specifically for the stock's currency and (in the case of beta) for the individual security.
Risk-Free Rate
We use the local 10-year sovereign bond yield, not a single global rate. Pulling a US Treasury rate to discount a German company's cash flows over-states their cost of equity by ~1.5–2 percentage points; using the Bund matches what local analysts do.
Market Risk Premium (MRP)
We use Damodaran's implied ERP series — the forward-looking expected premium derived from index dividends and growth — rather than the historical 5.5% US realised average. The implied series varies by region and is closer to what sell-side analysts use in their DCFs.
| Currency | MRP | Risk-Free Source |
|---|---|---|
| USD | 4.5% | 10Y Treasury |
| EUR | 4.2% | German Bund 10Y |
| GBP | 4.3% | UK Gilt 10Y |
| JPY | 5.0% | JGB 10Y |
| CHF | 4.0% | Swiss 10Y |
| CAD | 4.5% | Canada 10Y |
| AUD | 4.7% | Australia 10Y |
| INR | 7.5% | India 10Y |
| BRL | 8.5% | Brazil 10Y |
| CNY | 7.0% | China 10Y |
Refreshed annually from Damodaran's January implied-ERP update.
Beta — Bloomberg adjustment, no smoothing on top
The raw 5-year monthly beta from market data is one input; Bloomberg, Reuters, and most sell-side desks then apply a well-known adjustment before plugging into CAPM:
β_adjusted = (2/3) × β_raw + (1/3) × 1.0The reasoning: firm-level systematic risk mean-reverts toward the market over long horizons. A stock with β=2.5 today has historically shown its β drift toward 1.0 over 5–10 years. Discounting a 10-year DCF by raw β over-states the long-run equity-cost the firm is actually subject to.
We surface both numbers in the Forward DCF inputs panel: the adjusted β is what the model uses, with the raw value visible in the source label. We do notadditionally cap or smooth on top — beyond Bloomberg adjustment we trust the data. We only fall back to a sector median when the raw value is missing or implausible (β > 5, essentially always a Yahoo glitch on a thin ticker, not a real signal).
Sector medians (used as fallback only)
Sourced from Damodaran's industry data. Used when the underlying provider returned bad or missing data — never to replace a valid raw beta.
| Sector | Median β |
|---|---|
| Technology | 1.30 |
| Communication Services | 1.00 |
| Consumer Cyclical | 1.20 |
| Healthcare | 0.95 |
| Financial Services | 1.10 |
| Industrials | 1.05 |
| Real Estate | 0.95 |
| Energy | 0.90 |
| Materials | 1.05 |
| Consumer Defensive | 0.65 |
| Utilities | 0.55 |
Worked example
Siemens Energy (SIE.DE), EUR-listed, raw β = 1.81 from Yahoo:
β_adjusted = (2/3) × 1.81 + (1/3) × 1.0 = 1.54
Ke = 2.5% (German Bund 10Y)
+ 1.54 × 4.2% (EUR implied MRP)
= 2.5% + 6.5%
≈ 9.0%The stock's page shows this 11.0% in the "Cost of Equity (Ke)" row of the Forward DCF inputs panel, with each component sourced explicitly. If you disagree with any input — say you think β=1.81 is too high for a forward look — the slider on the valuation tab lets you override and recompute the fair value live.
Why we don't produce a single "official" Ke
Cost of equity is inherently an estimate. The Damodaran ERP we use is one of several reasonable choices; the 5y monthly beta is one of several reasonable beta horizons. We picked defensible defaults, surfaced them transparently, and gave you the tools to override. The number is a starting point for your analysis, not a verdict.
Last updated: 2026-05-02. Methodology subject to revision; check the page's git history at source.